RHEL6: Periodically RHEL NFS server no longer responds to NFS clients, requests to export cache timing out
Issue
- Periodically our RHEL6 NFS server stops responding to NFS client requests.
- On at least one occasion, when the problem occurred, someone was updating the /etc/exports file at the time with a script.
- In the messages file on the NFS server, we see nfsd: peername failed (err 107) errors, they occur 5 times, and then we know the NFS server is hung or soon to be hung.
- "Restarting" NFS clients seem to at least contribute, if not trigger the problem. By "restarting" we mean:
- "service portmap restart" and then "service autofs restart"
Environment
- Red Hat Enterprise Linux 6 (NFS server)
- seen on kernel 2.6.32-220.7.1.el6.x86_64, 2.6.32-431.el6.x86_64
- seen on nfs-utils-1.2.3-15.el6.x86_64 and nfs-utils-lib-1.1.5-4.el6.x86_64
- seen on nfs-utils-1.2.3-39.el6.x86_64 and nfs-utils-lib-1.1.5-6.el6.x86_64
- /etc/exports file contains thousands of lines
- NFSv4
- VMware
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.